CVE-2025-9905

The Keras Model.load_model method can be exploited to achieve arbitrary code execution, even with safe_mode=True. One can create a specially crafted .h5/.hdf5 model archive that, when loaded via Model.load_model, will trigger arbitrary code to be executed. This is achieved by crafting a special .h5 archive file that uses the Lambda layer feature of keras which allows arbitrary Python code in the form of pickled code. The vulnerability comes from the fact that the safe_mode=True option is not honored when reading .h5 archives. Note that the .h5/.hdf5 format is a legacy format supported by Keras 3 for backwards compatibility.

CVSS v3 7.3 HIGH

7.3^/10

CVSS v3 : HIGH

Vector :

Exploitability : 1.3 / Impact : 5.9

Attack Vector LOCAL

Attack Complexity LOW

Privileges Required LOW

User Interaction REQUIRED

Confidentiality Impact HIGH

Integrity Impact HIGH

Availability Impact HIGH

Scope UNCHANGED

References

Link	Resource
https://github.com/keras-team/keras/pull/21602	Issue Tracking Patch
https://github.com/keras-team/keras/security/advisories/GHSA-36rr-ww3j-vrjv	Exploit Vendor Advisory

Configurations

Configuration 1 (hide)

cpe:2.3:a:keras:keras:*:*:*:*:*:*:*:*

History

19 Oct 2026, 16:10

Type	Values Removed	Values Added
Summary		(es) El método Keras Model.load_model puede ser explotado para lograr ejecución de código arbitrario, incluso con safe_mode=True. Se puede crear un archivo de modelo .h5/.hdf5 especialmente diseñado que, al cargarse a través de Model.load_model, activará la ejecución de código arbitrario. Esto se logra al diseñar un archivo .h5 especial que utiliza la característica de capa Lambda de Keras, la cual permite código Python arbitrario en forma de código pickled. La vulnerabilidad surge del hecho de que la opción safe_mode=True no se respeta al leer archivos .h5. Tenga en cuenta que el formato .h5/.hdf5 es un formato heredado compatible con Keras 3 para compatibilidad con versiones anteriores.

Information

Published : 2025-09-19 09:15

Updated : 2026-10-19 16:10

NVD link : CVE-2025-9905

Mitre link : CVE-2025-9905

CVE.ORG link : CVE-2025-9905

JSON object : View

Products Affected

keras

keras

CWE

CWE-913

Improper Control of Dynamically-Managed Code Resources

{"id": "CVE-2025-9905", "cveTags": [], "metrics": {"ssvcV203": [{"source": "134c704f-9b21-4f2e-91b3-4a467353bcc0", "ssvcData": {"id": "CVE-2025-9905", "role": "CISA Coordinator", "options": [{"exploitation": "none"}, {"automatable": "no"}, {"technicalImpact": "total"}], "version": "2.0.3", "timestamp": "2025-09-20T03:55:41.389596Z"}}], "cvssMetricV31": [{"type": "Primary", "source": "nvd@nist.gov", "cvssData": {"scope": "UNCHANGED", "version": "3.1", "baseScore": 7.3, "attackVector": "LOCAL", "baseSeverity": "HIGH", "vectorString": "CVSS:3.1/AV:L/AC:L/PR:L/UI:R/S:U/C:H/I:H/A:H", "integrityImpact": "HIGH", "userInteraction": "REQUIRED", "attackComplexity": "LOW", "availabilityImpact": "HIGH", "privilegesRequired": "LOW", "confidentialityImpact": "HIGH"}, "impactScore": 5.9, "exploitabilityScore": 1.3}], "cvssMetricV40": [{"type": "Secondary", "source": "cve-coordination@google.com", "cvssData": {"Safety": "NOT_DEFINED", "version": "4.0", "Recovery": "NOT_DEFINED", "baseScore": 7.3, "Automatable": "NOT_DEFINED", "attackVector": "LOCAL", "baseSeverity": "HIGH", "valueDensity": "NOT_DEFINED", "vectorString": "CVSS:4.0/AV:L/AC:H/AT:P/PR:L/UI:P/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X", "exploitMaturity": "NOT_DEFINED", "providerUrgency": "NOT_DEFINED", "userInteraction": "PASSIVE", "attackComplexity": "HIGH", "attackRequirements": "PRESENT", "privilegesRequired": "LOW", "subIntegrityImpact": "HIGH", "vulnIntegrityImpact": "HIGH", "integrityRequirement": "NOT_DEFINED", "modifiedAttackVector": "NOT_DEFINED", "subAvailabilityImpact": "HIGH", "vulnAvailabilityImpact": "HIGH", "availabilityRequirement": "NOT_DEFINED", "modifiedUserInteraction": "NOT_DEFINED", "modifiedAttackComplexity": "NOT_DEFINED", "subConfidentialityImpact": "HIGH", "vulnConfidentialityImpact": "HIGH", "confidentialityRequirement": "NOT_DEFINED", "modifiedAttackRequirements": "NOT_DEFINED", "modifiedPrivilegesRequired": "NOT_DEFINED", "modifiedSubIntegrityImpact": "NOT_DEFINED", "modifiedVulnIntegrityImpact": "NOT_DEFINED", "vulnerabilityResponseEffort": "NOT_DEFINED", "modifiedSubAvailabilityImpact": "NOT_DEFINED", "modifiedVulnAvailabilityImpact": "NOT_DEFINED", "modifiedSubConfidentialityImpact": "NOT_DEFINED", "modifiedVulnConfidentialityImpact": "NOT_DEFINED"}}]}, "affected": [{"source": "cve-coordination@google.com", "affectedData": [{"repo": "https://github.com/keras-team/keras", "vendor": "Keras-team", "product": "Keras", "versions": [{"status": "affected", "version": "3.0.0", "versionType": "semver", "lessThanOrEqual": "3.11.2"}], "packageName": "keras", "collectionURL": "https://github.com/keras-team/keras", "defaultStatus": "unaffected"}]}], "published": "2025-09-19T09:15:36.033", "references": [{"url": "https://github.com/keras-team/keras/pull/21602", "tags": ["Issue Tracking", "Patch"], "source": "cve-coordination@google.com"}, {"url": "https://github.com/keras-team/keras/security/advisories/GHSA-36rr-ww3j-vrjv", "tags": ["Exploit", "Vendor Advisory"], "source": "cve-coordination@google.com"}], "vulnStatus": "Analyzed", "weaknesses": [{"type": "Secondary", "source": "cve-coordination@google.com", "description": [{"lang": "en", "value": "CWE-913"}]}], "descriptions": [{"lang": "en", "value": "The Keras Model.load_model\u00a0method can be exploited to achieve arbitrary code execution, even with safe_mode=True.\n\nOne can create a specially crafted .h5/.hdf5\u00a0model archive that, when loaded via Model.load_model, will trigger arbitrary code to be executed.\n\nThis is achieved by crafting a special .h5\u00a0archive file that uses the Lambda\u00a0layer feature of keras which allows arbitrary Python code in the form of pickled code. The vulnerability comes from the fact that the safe_mode=True\u00a0option is not honored when reading .h5\u00a0archives.\n\nNote that the .h5/.hdf5\u00a0format is a legacy format supported by Keras 3 for backwards compatibility."}, {"lang": "es", "value": "El m\u00e9todo Keras Model.load_model puede ser explotado para lograr ejecuci\u00f3n de c\u00f3digo arbitrario, incluso con safe_mode=True.\n\nSe puede crear un archivo de modelo .h5/.hdf5 especialmente dise\u00f1ado que, al cargarse a trav\u00e9s de Model.load_model, activar\u00e1 la ejecuci\u00f3n de c\u00f3digo arbitrario.\n\nEsto se logra al dise\u00f1ar un archivo .h5 especial que utiliza la caracter\u00edstica de capa Lambda de Keras, la cual permite c\u00f3digo Python arbitrario en forma de c\u00f3digo pickled. La vulnerabilidad surge del hecho de que la opci\u00f3n safe_mode=True no se respeta al leer archivos .h5.\n\nTenga en cuenta que el formato .h5/.hdf5 es un formato heredado compatible con Keras 3 para compatibilidad con versiones anteriores."}], "lastModified": "2026-10-19T16:10:00.260", "configurations": [{"nodes": [{"negate": false, "cpeMatch": [{"criteria": "cpe:2.3:a:keras:keras:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "315FD6FA-C1F3-47FD-AE5D-DF4D556D657F", "versionEndExcluding": "3.11.3", "versionStartIncluding": "3.0.0"}], "operator": "OR"}]}], "sourceIdentifier": "cve-coordination@google.com"}